Naive audience and communication bias

We introduce the possibility that the receiver naively believes the sender’s message in a game of information transmission with partially aligned objectives. We characterize an equilibrium in which the communication language is inflated, the action taken is biased, and the information transmitted is more precise than in the benchmark fully-strategic model. We provide comparative statics results with respect to the amount of asymmetric information, the proportion of naive receivers, and the size of the sender’s bias. As the state space grows unbounded, the equilibrium converges to the fully-revealing equilibrium that results in the limit case with unbounded state space.      

基于K-medoids聚类的贝叶斯集成算法

朴素贝叶斯分类算法由于其计算高效在生活中应用广泛。本文根据集成算法的差异性特征,聚类算法聚类点的选择方式的可变性,提出了基于K-medoids聚类技术的贝叶斯集成算法,朴素贝叶斯的泛化性能得到了提升。首先,通过样本集训练出多个朴素贝叶斯基分类器模型;然后,为了增大基分类器之间的差异性,利用K-medoids算法对基分类器在验证集上的预测结果进行聚类;最后,从每个聚类簇中选择泛化性能最佳的基分类器进行集成学习,最终结果由简单投票法得出。将该算法应用于UCI数据集,并与其他类似算法进行比较可得,本文提出的基于K-medoids聚类的贝叶斯集成算法(NBKME)提高了数据集的分类准确率。     

一种基于N-gram组合的中文垃圾邮件过滤方法

中文垃圾邮件的泛滥提出了极为迫切的技术诉求。本文使用了基于简单贝叶斯模型的过滤算法，同时使用N-gram对中文文本进行自动分词，并且组合多个N-gram来加快分类的收敛速度，这样分类是一种切实可行的垃圾邮件过滤方法。对于这种代价敏感性(cost sensitive)的分类，通过移动门限值的方法来处理：在评估结果时采用了TCR以及召回率(SR)和正确率(SP)等参数考察实验数据。实验表明：这种方法代价较小，而正确率较高。最后我们认为可以通过筛选训练邮件以及和其它措施相结合来满足ISP级别等应用场合的要求。     

基于读者视角的文本情感分类

从读者的角度对文本情感进行分类.训练样本集以新闻文章作为样本实例,以文章后读者的投票信息作为样本类别标注的先验知识.针对该不完备的数据集提出了一种半监督学习的分类模型,分类方法采用朴素贝叶斯分类法和EM算法相结合.实验证明该方法不仅简单有效,而且具有较高的分类性能.     

基于用户需求垃圾邮件过滤分类模型设计

以朴素贝叶斯理论作为基石并结合信息增益、代价因子等方法,尝试设计一种基于用户需求的垃圾邮件过滤分类模型,在垃圾邮件一次分类基础上提出邮件先过滤后分类方法,进而改进传统邮件分类一次阈值比较,变为两次阈值比较,且应用反馈技术以应对垃圾邮件的日益变化.该模型可能会对垃圾邮件多分类研究具有一定的参考价值.     

一种基于独立分量分析和Naive Bayesian网络的入侵检测方法

文章将独立分量分析(ICA)模型引入入侵检测系统，提出了基于独立分量分析和Naive Bayesian网络的入侵检测分类的新方法。通过把样本投影到有独立分量分析所确定的特征空间，来提高贝叶斯网络的分类性能，从而提高了入侵检测系统的性能。实验结果表明，这种基于独立分量分析模型的分类器具有良好的分类性能。     

Improved incremental algorithm of Naive Bayes

A novel Naive Bayes incremental algorithm was proposed,which could select new features.For the incremental sample selection of the unlabeled corpus,a minimum posterior probability was designed as the double threshold of sample selection by using the traditional class confidence.When new feature was detected in the corpus,it would be mapped into feature space,and then the corresponding classifier was updated.Thus this method played a very important role in class confidence threshold.Finally,it took advantage of the unlabeled and annotated corpus to validate improved incremental algorithm of Naive Bayes.The experimental results show that an improved incremental algorithm of Naive Bayes significantly outperforms traditonal incremental algorithm.     

Consumption and portfolio rules for time-inconsistent investors

This paper extends the classical consumption and portfolio rules model in continuous time [Merton, R.C., 1969. Lifetime portfolio selection under uncertainty: The continuous time case. Review of Economics and Statistics 51, 247–257, Merton, R.C., 1971. Optimum consumption and portfolio rules in a continuous time model. Journal of Economic Theory 3, 373–413] to the framework of decision-makers with time-inconsistent preferences. The model is solved for different utility functions for both, naive and sophisticated agents, and the results are compared. In order to solve the problem for sophisticated agents, we derive a modified HJB (Hamilton–Jacobi–Bellman) equation. It is illustrated how for CRRA functions within the family of HARA functions (logarithmic and power utilities) the optimal portfolio rule does not depend on the discount rate, but this is not the case for a general utility function, such as the exponential (CARA) utility function.     

基于PLS的加权朴素贝叶斯分类测试算法

朴素贝叶斯算法是一种简单而高效的分类算法,但是它的条件独立性假设影响了其分类性能。通过放松朴素贝叶斯假设,可以增强其分类效果,但通常会导致计算代价大幅提高。文章提出了一种基于偏最小二乘的加权朴素贝叶斯分类算法,通过建立条件属性和决策属性之间偏最小二乘回归方程,把回归系数赋给对应的条件属性,作为相应的权重,从而在保持简单性的基础上有效地提高了朴素贝叶斯算法的分类性能。最后,通过在UCI数据集上的仿真实验,验证了该算法的有效性。